A publication of the James F. Lincoln Arc Welding Foundation

Volume XIX, Number 1, 2002
A publication of the James F. Lincoln Arc Welding Foundation

Continuing the Tradition…
It was with humility that I recently accepted
Duane Miller’s offer to assist in the editing of
Welding Innovation. This was true because I did
not think it would be possible for me to fill the
position left vacant by my predecessor Scott
Funderburk. He covered Welding Innovation
for several years as assistant editor, and he
mentored well under Duane’s watchful eye. I
would expect that Scott will be sorely missed
here in the application engineering department
and by our readership. We all wish him the very
best in his new assignment.
As is certainly the case with the roles of most
professionals employed in industry today, we are
required to wear many hats, and my primary
position as senior application engineer has
dovetailed nicely with my additional responsibility
to Welding Innovation. The irony of this is that
nearly twenty years ago, I was asked to assist
the former editor, Richard Sabo, by suggesting a
name for this publication. We finally settled on
The Welding Innovation Quarterly, and that
name became synonymous with high-quality,
creative design engineering concepts coupled
with fundamentally strong welding principles. If
you needed answers to tough problems, then in
return we were prepared as a group to provide
high quality information. In principle, it is our
objective to continue the rich tradition of
Welding Innovation, and I would like to suggest
that the articles in this issue of the magazine
reflect our commitment to that objective.
Professor Henry Petroski’s contribution to this
issue, “The Fall of Skyscrapers,” provides an
analysis of the collapse of the World Trade
Center towers resulting from the September 11
terrorist attack. In addition, Professor Petroski
provides us with ideas regarding future design
considerations, and the practicality of future
skyscraper construction. In contrast, contributing
writer Carla Rautenberg provides an article that
presents Frank Gehry’s free-form design of the
Peter B. Lewis Building at the Weatherhead
School of Management located on the Case
Western Reserve University campus. The compelling
design of the newly constructed Gehry
design is a clear break from the historical
constraints of rectangular designed structures.
We are confident that this issue of Welding
Innovation will be thought-provoking, and we
encourage your innovative input for future issues.
I especially urge you to consider sharing with our
readers the value of your own experience, in the
form of a submission to the “Lessons Learned in
the Field” column initiated by our senior design
consultant, Omer W. Blodgett. See page 18 for
another of Omer’s timeless and valuable engineering
lessons.
Jeff Nadzam
Assistant Editor
Australia and New Zealand
Raymond K. Ryan
Phone: 61-2-4862-3839
Fax: 61-2-4862-3840
INTERNATIONAL SECRETARIES
Croatia
Prof. Dr. Slobodan Kralj
Phone: 385-1-61-68-222
Fax: 385-1-61-56-940
Russia
Dr. Vladimir P. Yatsenko
Phone: 077-095-737-62-83
Fax: 077-093-737-62-87

Cover: With a shimmering stainless steel
skin concealing its intricate welded structure,
the roof of the almost-completed Peter B.
Lewis Building at Case Western Reserve
University definitely makes an architectural
statement. See story on page 20.
Omer W. Blodgett, Sc.D., P.E.
Design Consultant
Volume XIX
Number 1, 2002
Editor
Duane K. Miller,
Sc.D., P.E.
Assistant Editor
Jeff R. Nadzam
The James F. Lincoln
Arc Welding Foundation
The views and opinions
expressed in Welding
Innovation do not necessarily
represent those of
The James F. Lincoln Arc
Welding Foundation or The
Lincoln Electric Company.
The serviceability of a
product or structure utilizing
the type of information
presented herein is, and
must be, the sole responsibility
of the builder/user.
Many variables beyond the
control of The James F.
Lincoln Arc Welding
Foundation or The Lincoln
Electric Company affect the
results obtained in applying
this type of information.
These variables include,
but are not limited to, welding
procedure, plate chemistry
and temperature,
weldment design, fabrication
methods, and service
requirements.
Features
2 Orthotropic Design Meets Cold Weather Challenges
This overview of orthotropic steel deck bridges discusses a number
of examples that have been built in Norway, Russia, Sweden and Ukraine.
12 The Fall of Skyscrapers
In an article reprinted from American Scientist, Duke University Professor
Henry Petroski considers some ramifications of the collapse of the World Trade
Center towers on the future skylines of the world’s cities.
20 Gleaming “Waterfall” Refreshes Urban Campus
The work of architect Frank Gehry inspires controversy and excitement
as the new home of Case Western Reserve University’s Weatherhead School
nears completion in Cleveland, Ohio.
Departments
7 Design File: Designing Fillet Welds for Skewed T-Joints, Part 1
11 Opportunities: Lincoln Electric Technical Programs
17 Opportunities: Lincoln Electric Professional Programs
18 Lessons Learned in the Field: Consider the Transfer of Stress
through Members
Visit us online at www.WeldingInnovation.com
THE JAMES F. LINCOLN ARC WELDING FOUNDATION
Dr. Donald N. Zwiep, Chairman
Orange City, Iowa
John Twyble, Trustee
Mosman, NSW, Australia
Duane K. Miller, Sc.D., P.E.
Executive Director & Trustee
Roy L. Morrow
President
Welding Innovation Vol. XIX, No. 1, 2002
1

Orthotropic Design Meets Cold Weather Challenges
An Overview of Orthotropic Steel Deck Bridges in Cold Regions
By Alfred R. Mangus, P.E.
Transportation Engineer, Civil
California Dept. of Transportation (CALTRANS)
Sacramento, California
Introduction
Initially developed by German engineers
following World War II, orthotropic
bridge design was a creative response
to material shortages during the postwar
period. Lightweight orthotropic steel
bridge decks not only offered excellent
structural characteristics, but were also
economical to build (Troitsky 1987).
Moreover, they could be built in cold
climates at any time of the year.
Engineers from around the world utilize
this practical and economic system for
all types of bridges. While concrete
must be at or above 5 degrees Celsius
to properly cure, it is physically possible
to encapsulate and heat the concrete
construction process; admittedly, this
adds to construction costs (Mangus
1988), (Mangus 1991).
Orthotropic steel deck bridges have
proven to be durable in cold regions.
The orthotropic steel deck integrates
the driving surface as part of the
superstructure, and has the lowest
total mass of any practicable system.
In Europe, where the advantages of
orthotropic design have been
embraced with enthusiasm, there are
more than 1,000 orthotropic steel deck
bridges. In all of North America, there
are fewer than 100 bridges of
orthotropic design.
This article will give an overview of
imaginative steel deck bridges currently
in operation in Norway, Russia,
Sweden, and Ukraine. The examples
Figure 1. Rib designs.
cover a matrix of rib types, superstructure
types and various bridge types.
Russia has developed a mass manufactured
panelized orthotropic deck
system and has devised special
launching methods for cold regions.
Russian engineers prefer the open rib
design and have industrialized this
system, while most other engineers
prefer the closed rib. Researchers, as
well as the owners of orthotropic steel
deck bridges, continue to monitor the
performance of various rib types
(Figure 1).
In Norway
Nordhordland Floating Bridge
The Nordhordland Bridge across the
Salhus fjord is Norway’s second floating
bridge and the world’s largest floating
bridge (Meaas, Lande, and Vindoy,
1994). The bridge was opened for traffic
in 1994. The total bridge length is
1615 m and consists of a high level
369 m long cable stayed bridge and a
1246 m long floating bridge (Figure 2).
The floating bridge consists of a steel
box girder, which is supported on ten
concrete pontoons and connected to
abutments with transition elements in
forged steel. The main elements are a
high-level cable stayed bridge providing
a ship channel and a floating
bridge between the underwater rock
Klauvaskallen and the other side of
the fjord. The cable stayed bridge provides
a clear ship channel. A 350 m
long ramp is required to transition from
the higher bridge deck on the cable
2 Welding Innovation Vol. XIX, No. 1, 2002
Figure 2. Nordhordland Floating Bridge across Salhus fjord of Norway.

stayed bridge to the bridge deck 11 m
above the waterline. The steel box
girder of the floating bridge forms a
circular arch with a radius of 1,700 m
in the horizontal plane. The supporting
pontoons are positioned with a center
distance of 113 m and act as elastic
supports for the girder, which is
designed without internal hinges. The
bridge follows the tidal variations by
elastic deformations of the girder.
The steel box girder is the main loadcarrying
element of the bridge (Figure
3). The octagon girder is 5.5 m high
and 13 m wide. The free height below
the girder down to the waterline is 5.5
m and this allows for passage of small
boats. The plate thickness varies from
14 mm to 20 mm. The plate stiffeners
are in the traditional trapezoidal shape
and they span in the longitudinal direction
of the girder. The stiffeners are
supported by cross-frames with center
distance of maximum 4.5 m. At the
supports on the pontoons, bulkheads
are used instead of cross-frames. This
is done because the loads in these
sections are significantly larger than in
the cross-frames. The plate thickness
Figure 3. Nordhordland Floating
Bridge.
in the bulkheads varies from 8 mm to
50 mm. The box girder is constructed
in straight elements with lengths varying
from 35 m to 42 m. The elements
are welded together with a skew angle
of 1.2° to 1.3° to accommodate the
arch curvature in the horizontal plane.
The cross section dimensions of the
octagonal box girder are constant for
the length of the bridge.
The stress level varies significantly
over the length of the bridge. In the
areas with the highest stresses, steel
with a yield strength of 540 MPa is
used.
Orthotropic bridge design
was a creative response
to material shortages
during the postwar period
In the remainder of the bridge (in the
cross-frames and bulkheads) normalized
steel with yield strength of 355
MPa is used. The total steel weight of
the box girder is 12,500 tons, of which
approximately 3,000 tons are high
strength steel. The elevated ramp is
approximately 350 m long and has a
grade of 5.7 percent (Figure 4). The
elevated ramp is constructed with an
orthotropic plate deck 12 mm thick
and has 8 mm and 10 mm thick trapezoidal
ribs 800 mm deep. T-shaped
crossbeams support the ribs with a
maximum center distance of 4.5 m.
The main 1,200 mm deep box girders
are located one at each edge of the
ramp in order to maximize the stiffness
about a vertical axis. The steel weight
of the ramp is 1,600 tons.
Figure 4. Nordhordland Floating
Bridge.
Bybrua Bridge
The Bybrua cable stayed bridge has a
main span of 185 m. The 15.5 m wide
roadway superstructure was fabricated
in the shop in 9.0 m sections (Figure 5).
There is a combined pedestrian plus
bicycles area on each side of the three
traffic lanes. The cross section of the
main span has a deck-plate 12 mm
thick, but this increases to 16 and 20
mm at the cable anchorage. The bottom
plate varies between 8 mm and
10 mm thickness, and the webs
between 12 mm and 20 mm. At
intervals of 3.0 m there are frames
supporting the longitudinal stiffening
system. In the bridge deck this is
made up of standard trapezoidal ribs
Figure 5. Bybrua Bridge.
Welding Innovation Vol. XIX, No. 1, 2002 3

from the German steel company
Krupp, and in the bottom flange box
section of bulb flats open ribs were utilized
(Aune and Holand 1981). In the
longitudinal direction the deck was
divided into six fabricated sections,
two of which were welded to the web
sections. The box bottom was fabricated
as three sections. The total steel
weight is about 1,100 tons. All elements
prefabricated in the shop were
welded, as were the field splices in the
deck, whereas “Huck” high tensile
bolts were used in all other field joints.
All field joints were calculated as friction
connections. The whole steel
structure is metallized with zinc and
painted according to the specifications
of the Norwegian Public Roads
Administration. The superstructure
received the maximum live load
stresses during the erection of the
bridge. The wearing surface of the
bridge deck is the same as that developed
by the Danish State Road
Laboratories for the Lillebelt Bridge of
Denmark.
Storda and Bomla Bridges
The “Triangle Link” project connects
three islands off the Norwegian coast
south of Belgen to the mainland with
three bridges (Larson and Valen
Researchers continue
to monitor the performance
of various rib types
2000). The entire project was completed
in April 2001. The two orthotropic
steel deck suspension bridges are
known as the Storda Bridge and the
Bomla Bridge. The Storda Bridge is
1,076 m long, has a main span of 677
m, with towers 97 m high and a vertical
clearance of 18 m (Figure 6). The
Bomla Bridge is 990 m long with a
main span of 577 m and the tower
height is 105 m. The roadway of both
bridges is 9.7 m wide. Scanbridge AS
of Norway fabricated the Bomla
Bridge’s steel approach superstructure,
which was launched out over the
tops of the columns from the shore.
The steel components for the main
span superstructure of the Storda
Bridge were prefabricated in the
Netherlands (Figure 6) and the main
span superstructure of the Bomla
Bridge was prefabricated in Italy. The
orthotropic ribs for the Storda Bridge
were prefabricated in France. The
orthotropic sections were transported
to the site by barge, and were lifted
into position by a crane.
Figure 6. Storda Bridge.
In Russia
Russian engineers have standardized
their orthotropic deck plates using
open or flat plate ribs as shown in
Figure 1. They have several launching
solutions or standardized methods for
pushing the superstructure across a
river or gorge. There are a limited
number of bridge case histories documented
in English, but they provide an
overall view of Russian techniques
(Blank, Popov, and Seliverstov 1999).
In the city of Arkhangel, Russia, a vertical
lift record span bridge of 120.45
m was completed in 1990 (Stepanov
1991). The Berezhkovsky twin parallel
bridges are multi-cell box girder
bridges consisting of three spans of
110 m + 144.5 m +110 m. Each bridge
has four traffic lanes 3.75 m wide.
These bridges were the first to be
launched with inclined webs
(Surovtsev, Pimenov, Seliverstov, and
Iourkine 2000).
Oka Bridge
The four-lane orthotropic twin box
girder bridge crossing the Oka River
on the bypass freeway around the city
of Gorki, Russia, was opened to traffic
in 1991 (Figure 7). The 966 m long
superstructure consists of 2 spans x
84 m + 5 spans x 126 m + 2 spans x
84 m (Design Institute Giprotransmost
1991). This bridge is a single continuous
superstructure with a fixed bearing
420 m away from one of the abutments.
The total bridge width (29.5 m
including steel traffic barriers) provides
two sidewalks 1.5 m wide each, four
traffic lanes, four safety shoulders and
a center median. The total weight of
steel for the superstructure is 10,635
tons, or 373 kg/m2. The orthotropic
steel superstructure comprises five
basic elements (Figure 7). There are
two main box girders assembled from
two L-shaped sections for the bottom
face and sides. The intermediate
orthotropic plate sections were used
for the top flange of the two box girders,
as well as the majority of the
deck. The end sections of the
orthotropic plate were panelized with
tapered ends, because only sidewalk
loading is required. The transverse
diaphragms are steel trusses between
the box girders. The diaphragms
required extra steel beams at the bottom
flange of the box girder above the
bearings. The main box girder was
shop fabricated in L-shaped sections
that are 21 m long and 3.6 m deep.
The intermediate orthotropic welded
steel deck plate was shop fabricated in
panelized sections 2.5 m wide and
11.5 m long.
The longitudinal ribs of the orthotropic
deck and steel box girders are flat rib
plates spaced at 0.35 m, and the
spacing of transverse ribs is 3.0 m for
both components. The stiffening ribs of
the main girder are located on both
sides of every web. The vertical split-T
ribs of the box girders were aligned
with the transverse ribs of the ribbed
plate, thus creating the integral internal
diaphragms. The longitudinal stiffening
ribs are at a constant spacing
4 Welding Innovation Vol. XIX, No. 1, 2002

Figure 7. Twin box girder bridge crossing the Oka River, Russia (split–section).
along the bridge. Depending on the
web thickness, additional vertical stiffening
ribs were required between the
diaphragms. The superstructure was
erected using “continuous launching”
from one bank of the Oka River. The
shop-fabricated elements were added
piece by piece to form a continuous
structure at the “erection slip” area on
this riverbank. The joints of the horizontal
sections of the orthotropic deck
and ribbed plates, as well as the joints
of the web of the main girder, were
automatically welded. The joints of the
longitudinal ribs of the ribbed plate
were manually welded. All the remaining
joints used high strength bolts.
In Sweden
High Coast Suspension Bridge
The High Coast Suspension Bridge of
Sweden is almost the same size as
San Francisco’s Golden Gate Bridge
(Merging with Nature 1998). The main
span is 1,210 m long with suspended
abutment to abutment where expansion
joints and hydraulic buffers are
located. The 48 box girder sections
were fabricated at a shipyard in
Finland (Pedersen 1997). The standard
section is 20 m long with two
Figure 9. High Coast, Sweden
components.
sets of hangers each and weighs 320
tons (Figure 9). The 20 m long panels
for the deck, sides and bottom were
fabricated with a maximum width of 10
m. They were fabricated from steel
plates, typically 9 to 14 mm thick,10 m
long and 3 to 3.3 m wide. The ribs
were 20 m long trapezoidal ribs with a
plate thickness of 6 to 8 mm.
The plates were placed on a plane
and welded in the transverse and longitudinal
direction and the trough stiffeners
were fitted and welded
longitudinally. Plates connecting the
panels and the diaphragms were welded
between ribs. The 20 m long edge
sections and units for the transverse
diaphragm, or bulkhead, were prefabricated.
The bottom and inclined sides
were placed first. Each 4 m deep
transverse diaphragm or bulkhead was
fitted. The edge sections were installed
and finally the two deck panels were
placed on top, completing a 20 m long
subsection. The 31 bridge girder sections
for the main span were transported
from the fabrication yard in Finland
on the three barges in the same way
as the sections for the side spans, and
erected with a floating sheerleg crane,
130 m boom, starting from mid-span
and proceeding towards the towers
(Edwards and Westergren 1999).
In Ukraine
South Bridge over the Dnipro River
The 1992 signature span of the South
Bridge over the Dnipro [Dnepr] River
in Kyiv [Kiev], Ukraine is an unsymmetrical
cable stayed bridge with a
main span of 271 m (Korniyiv and.
Fuks 1994). The main span side of the
H-tower is a continuous three-span
steel box girder with orthotropic steel
deck. The back-span superstructure on
the opposite side of the H-tower is a
segmental prestressed concrete box
section. Concrete construction for the
shorter back span of 60 m was used
as a counter-weight mass equal to the
longer orthotropic main span. The
bridge carries a six-lane roadway, two
rail tracks and four large-diameter
water pipes (Figure 10). The total live
Figure 8. High Coast, Sweden, section.
side spans of 310 m and 280 m. The
width of the roadway is 17.8 m, allowing
for a possible future extension to 4
lanes (Figure 8). The distance
between the main cables is 20.8 m
and there are 20 m between the hangers.
The girder is continuous through
the towers extending 1,800 m from
Figure 10. South Bridge (cable stayed) of Ukraine (split–section).
Welding Innovation Vol. XIX, No. 1, 2002 5

load is about 246 kN/m. The threespan
(80.5 m + 90 m + 271 m) continuous
steel box girders are made of
low-alloy steel with a minimum yield
strength of 390 MPa.
The bridge was divided into segments
that were shop-welded. Field splices
were either welded or joined by highstrength
bolts. Bolting was used where
automatic welding was impractical
because of the short length of the
weld or difficult access. The cross section
of the twin two-cell box girder
bridge consists of six vertical webs,
the upper deck plate and the lower
box flanges. The narrow cell is for the
cable stayed bridge anchorage. In the
central portion of the cross section
that carries the two hot water supply
pipes, the lower flange was omitted to
preclude the undesirable effects of
unequal heating inside a closed box.
The bearings at the piers permit lateral
displacement of the superstructure
because of the 41.5 m bridge width.
The orthotropic decks, bottom flanges
of the boxes and the webs have open
flat-bar stiffening ribs, a common feature
in Ukrainian and Russian bridges.
Longitudinal flat or open ribs were
placed on the top face of the
orthotropic deck plate under the rail
tracks, thus avoiding intersections of
longitudinal and transverse stiffeners.
This facilitated fabrication, at the same
time precluding stress concentrations
at crossing welds that would be susceptible
to fatigue under dynamic train
loading. Longitudinal ribs under the
train tracks have a depth in excess of
the design requirements, which permitted
longitudinal profile adjustments
of the tracks after erection of the
bridge superstructure. The steel girders
were pre-assembled on the bank
of the south and erected by launching.
The twin two-cell box girders were
equipped with a launching nose and
stiffened with a temporary strut system
(Rosignoli1999). Single erection rollers
were used at the tops of supports and
had a friction factor of less than 0.015.
The erection of the 271 m main span
was accomplished with two false work
supports providing three equal spans
of about 90 m. At the H-tower of the
cable stayed bridge, hinged conditions
are provided by supports with limited
rotational capability in the vertical and
the horizontal planes. The torsional
rigidity of the bridge is supplied by the
two planes of cable stays, plus the
stiffness of the closed box sections.
Under one-sided loading of the bridge
(three traffic lanes), the deck cross
slope of 0.3% was measured in field
tests, less than the calculated value of
0.35%. Eccentric hinged connections
between the bottom flanges of the
stiffening girders and the H-tower were
constructed, considerably reducing the
bending moments in the girders.
Conclusion
The foregoing examples illustrate a
range of creative responses to the
challenge of designing and constructing
orthotropic steel deck bridges in
cold weather regions. The versatility,
economy and structural integrity of
welded orthotropic design undoubtedly
will continue to inspire bridge designers
and structural engineers in the
21st century.
Figure Credits
Figure 1 from Ballio, G., Mazzolani, F. M. 1983,
“Theory and Design of Steel Design Structures,”
Chapman & Hall Ltd. courtesy of Dr. Mazzolani;
Figures 2, 3, & 4 courtesy of Dr. Ing. A-
Aas–Jakobsen, AS Structural Engineering
Consultants, Oslo, Norway; Figure 5 adapted from
Aune, Petter, and Holand, Ivar (1981); Figure 6
courtesy of Mr. L. Adelaide of Profilafroid, France;
Figures 8 & 9 courtesy of Claus Pedersen of
Mondberg & Thorsen AS, Copenhagen, Denmark;
Figures 7 & 10 courtesy of IABSE International
Association of Bridge Structural Engineers.
References
Aune, P., Holand, I. (1981) Norwegian Bridge
Building – A Volume Honoring Arne Selberg,
Tapir, Norway
Blank, S. A., Popov, O. A., Seliverstov, V. A., (1999)
“Chapter 66 Design Practice in Russia,” Bridge
Engineering Handbook, 1st ed., Chen, W.F.,
and Duan, L. Editors, CRC Press, Boca Raton
Florida
Design Institute Giprotransmost, (1991) “The
Bridge-Crossing over the River Oka on
the Bypass Road near Gorki” Structural
Engineering International, IABSE, Zurich,
Switzerland, Vol. 1, Number 1, 14-15
Edwards,Y. & Westergren P., (1999), “Polymer
modified waterproofing and pavement system
for the High Coast Sweden” NordicRoad &
Transport Research No. 2, 28
Korniyiv, M. H. and Fuks, G. B (1994) “The South
Bridge, Kyiv, Ukraine” Structural Engineering
International, IABSE, Zurich, Switzerland, Vol.
4, Number 4, 223-225
Larsen J., and Valen, A., (2000) “Comparison of
Aerial Spinning versus Locked-Coil Cables
for Two Suspension Bridges (Norway),”
Structural Engineering International, IABSE,
10(3), 128-131
Mangus, A., (1988) “Air Structure Protection of
Cold Weather Concrete,” Concrete
International, American Concrete Institute,
Detroit, MI, October 22-27
Mangus, A., (1991) “Construction Activities Inside
of Air Structures Protected From the Arctic
Environment,” International Arctic Technology
Conference, Society of Petroleum Engineers,
Anchorage, AK, May 29-31
Mangus, A, and Shawn, S., (1999) “Chapter 14
Orthotropic Deck Bridges,” Bridge Engineering
Handbook, 1st ed., Chen, W.F., and Duan, L.
Editors, CRC Press, Boca Raton Florida
Mangus, A., (2000) “Existing Movable Bridges
Utilizing Orthotropic Bridge Decks,” 8th Heavy
Movable Bridge Symposium Lake Buena Vista
Florida, Heavy Movable Structures, P. O. Box
398, Middletown, NJ
Merging with Nature – Hoga Kusten [High Coast]
(1998), Bridge Design & Engineering, 10, (1),
50–56
Meaas, P., Lande, E., Vindoy, V., (1994) “Design of
the Salhus Floating Bridge (Nordhordland
Norway),” Strait Crossings 94, Balkema,
Rotterdam ISBN 90-5410 388-4 3(3), 729-734
Pedersen, C., (1997) The Hoga Kusten (High
Coast) Bridge-suspension bridge with 1210 m
main span –construction of the superstructure,
Mondberg & Thorsen A/S Copenhagen,
Denmark
Rosignoli, M.,(1999) Launched Bridges –
Prestressed Concrete Bridges built on the
ground and launched into their final position,
ASCE Press, Reston, VA
Stepanov, G. M., (1991), “Design of Movable
Bridges,” Structural Engineering International,
IABSE, Zurich, Switzerland, Volume 1,
Number 1, 9-1
Surovtsev, V., Pimenov, S., Seliverstov, V., &
Iourkine, S, (2000) “Development of Structural
Forms and Analysis of Steel Box Girders with
inclined webs for operation and erection conditions”
Giprotransmost J.S.Co, Pavla
Kortchagina str. 2, 129278 Moscow, Russian
Federation
Troitsky, M. S., (1987) Orthotropic Bridges - Theory
and Design, 2nd ed., The James F. Lincoln Arc
Welding Foundation, Cleveland, OH
6 Welding Innovation Vol. XIX, No. 1, 2002

Design File
Designing Fillet Welds for Skewed T-joints—Part 1
Practical Ideas for the Design Professional by Duane K. Miller, Sc.D., P.E.
Introduction
Detailing fillet welds for 90-degree T-joints is a fairly
straightforward activity. Take the 90-degree T-joint and skew
it—that is, rotate the upright member so as to create an
acute and obtuse orientation, and the resultant geometry
of the fillet welds becomes more complicated (see Figure
1). The greater the degree of rotation, the greater the difference
as compared to the 90-degree counterpart.
A series of equations can be used to determine weld sizes
for various angular orientations and required throat dimensions.
Since the weld sizes on either side of the joint are
not necessarily required to be of the same size, there are a
variety of combinations that can be used to transfer the
loads across the joint. While there are theoretical savings
to be seen by optimizing the combinations of weld sizes,
rarely do such efforts result in a change in fillet weld size of
even one standard size.
Codes prescribe different methods of indicating the
required weld size. These are summarized herein.
When acute angles become smaller, the difficulty of
achieving a quality weld in the root increases. The AWS
D1.1 Structural Welding Code deals with this issue by
requiring the consideration of a Z-loss factor.
This edition of Design File addresses the situation where
the end of the upright member in the skewed T-joint is parallel
to the surface of the other member. A future Design
File column will consider the situation in which the upright
member has a square cut on the end, resulting in a gap on
the obtuse side. Also to be addressed in the future are
weld options other than fillet welds in skewed T-joints.
LEG
SIZE
DIHEDRAL ANGLE
0
135 MAX
ω 2
ω 1
ψ 2
ω ψ 1
1
f 2
ω 2
b 2 b 1
f 1
t 2
t 1
b 2
b 1
Figure 1. Equal throat sizes (t 1 = t 2 ).
DIHEDRAL
ANGLE
0
60 MIN
The Geometry
Figure 1 provides a visual representation of the issue. For
the 90-degree orientation, the weld throat is 70.7% of the
weld leg dimension. This relationship does not hold true for
fillet welds in skewed joints. On the obtuse side, the weld
throat is smaller than what would be expected for a fillet
weld of a similar leg size in a 90-degree joint, and the
opposite is the case for the acute side. These factors must
be considered when the fillet weld leg size is determined
and specified.
Careful examination of the fillet welds on the skewed joint
raises this question: What is the size of the fillet weld in a
skewed joint?
LEG
SIZE
Figure 1 illustrates the fillet weld leg size for a skewed T-
joint, and is designated by “ω.” This, however, is inconsistent
with AWS Terms and Definitions (AWS A3.0-94) which
defines a “fillet weld leg” as “The distance from the joint
root to the toe of the fillet weld.” According to this definition,
and as shown in Figure 1, the fillet weld leg is dimension
“b.” The dimension that is labeled “ω” is the distance from a
member to a parallel line extended from the bottom weld
Welding Innovation Vol. XIX, No. 1, 2002 7

toe. While not technically correct according to AWS A3.0, it
is the dimension and terminology used when fillet welds in
skewed joints are discussed in the AWS D1.1 Structural
Welding Code, as well as other AWS publications (i.e., The
Welding Handbook, ninth edition, volume 1). Such terminology
will be used here.
This raises an additional question: What would a weld
inspector actually measure when dealing with a fillet weld
in a skewed T-joint? Conventional fillet weld gauges could
be used to measure the obtuse side’s fillet weld leg dimension
“ω” as shown in Figure 1. Dimension “b” would be difficult
to measure directly since the location of the weld root
cannot be easily determined. Welds on the acute side are
impossible to measure using conventional fillet weld
gauges. The face dimension “f,” however, offers an easy
alternative: when this dimension is known for the weld size
and the dihedral angle, the welder and inspector can easily
determine what the actual size is by using a pair of
dividers. Alternately, a series of simple gauges of various
widths could be made to directly compare the requirements
to the actual weld size. Thus, dimension “f” may be important
for controlling weld sizes in skewed T-joints.
When sizing a fillet weld for 90-degree T-joints or skewed T-
joints, the starting point is to determine the required throat
size needed to resist the applied loads. From the throat
dimension, the fillet weld leg size can be determined. Three
options will be considered:
Where the throat size is the same on either side of the
joint (see Figure 1)
To determine the required fillet weld size for a given throat,
the following relationship can be used:
ψ
ω = 2 t sin
(
2
) (1)
The width of the face of the weld (“f”) can be found from
this equation:
(2)
f = 2 t tan (
ψ
2 )
Dimension “b,” that is, the ‘true’ fillet leg size, can be found
from this relationship: t
b =
cos (
ψ
2 )
(3)
Finally, the cross-sectional area of the weld metal can be
determined from the following:
A = t 2 tan (
ψ
2 )
(4)
Where the leg size is the same on both sides
(see Figure 2)
If the designer decides to make both welds with the same
leg size (as is illustrated in figure 2), the first required step
is to determine the composite total dimension of the two
8 Welding Innovation Vol. XIX, No. 1, 2002
ω 2
b
ω 1
b
b
t 2
Figure 2. Equal fillet weld leg sizes (ω 1 = ω 2 ).
throat sizes. This dimension “t T ” is then inserted into the following
equations to determine the two throats “t 1 ” and “t 2 .”
ψ
cos
t 1
= t T
(
1
2 )
ψ
cos (
1
)
ψ
+ cos (
2
)
2
ψ
cos
t 2
= t T
(
2
2 )
ψ
cos (
1
)
ψ
+ cos (
2
)
2
Equations 1 – 4 can be used to find the corresponding fillet
weld leg size, face dimension, “b” dimension, and crosssectional
area. These calculations will be made using the
applicable throat dimension “t” determined from equations
5 and 6, not the total throat dimension “t T ” used in equations
5 and 6.
Where a minimum quantity of weld metal is used
(see Figure 3)
Even a casual review of Figure 1 shows that, when fillet
weld leg sizes are specified to be of the same size on
either side of the skewed T-joint, the use of weld metal is
as efficient as could be. Minimum weld metal can be
obtained by taking advantage of the more favorable condition
that results on the acute side where a greater weld
throat can be obtained for the same quantity of metal that
would be placed on the obtuse side.
To minimize the volume of weld metal used in the combination
of the two welds, the following equations may be used
once the total throat dimension “t T ” is known:
t 1
=
t 2
=
1 + tan 2 (
1 + tan 2 (
t T
ψ1
)
t T
2
ψ2
)
2
t 1
b
2
2
(5)
(6)
(7)
(8)

Although the preceding calculations are not particularly difficult,
Table 1 has been provided to simplify the process.
Columns A and B are used to determine fillet weld leg
sizes and face widths for various dihedral angles. To obtain
the required fillet weld size, the calculated throat is multiplied
by the factor in Column A. Face widths can be found
following the same procedure.
t 2
b 2
b 1
t 1
If the same leg size is desired on either side of the joint,
columns C-E are used. In this case the total weld throat “t T ”
is used, as opposed to what was done with columns
A and B.
For the minimum weld volume, columns F–H can be used.
Again, the total weld throat “t T ” is used.
As will be discussed below, for dihedral angles of 30–60
degrees, the D1.1 Code requries the application of a Z-loss
factor. Thus, the values in Table 1 that apply to dihedral angles
where this applies are shown in blue numbers to remind the
user to incorporate this factor into the weld throat sizes.
b 2
b 1
Figure 3. Minimum weld volume.
Influence of Dihedral Angle
AWS D1.1 Structural Welding Code–Steel provides for five
groupings of skewed T-joints, depending on the range of
sizes of the dihedral angle: a) Obtuse angles greater than
100 degrees, b) angles of 80–100 degrees, c) acute angles
of 60–80 degrees, d) acute angles of 30–60 degrees, and
e) acute angles of less than 30 degrees. Each is dealt with
in a slightly different manner.
Dihedral Angle
Leg & Face Dimensions
Multipy by t
Table 1.
Same Leg Size
Multipy by t T
Minimum Weld Volume
Multipy by t T
Ψ A B C D E F G H
phi1 deg leg size face width throat leg size face width throat leg size face width
30 0.517 0.536 0.788 0.408 0.422 0.933 0.483 0.536
35 0.601 0.630 0.760 0.457 0.479 0.910 0.547 0.630
40 0.684 0.728 0.733 0.501 0.533 0.883 0.604 0.728
45 0.765 0.828 0.707 0.541 0.585 0.854 0.653 0.828
50 0.845 0.932 0.682 0.576 0.635 0.822 0.694 0.932
55 0.923 1.04 0.657 0.607 0.684 0.787 0.726 1.04
60 1.00 1.15 0.634 0.634 0.731 0.750 0.750 1.15
65 1.07 1.27 0.611 0.656 0.778 0.712 0.764 1.27
70 1.15 1.40 0.588 0.674 0.823 0.671 0.770 1.40
75 1.22 1.53 0.566 0.689 0.868 0.630 0.766 1.53
80 1.29 1.68 0.544 0.699 0.912 0.587 0.755 1.68
85 1.35 1.83 0.522 0.705 0.956 0.544 0.735 1.83
90 1.41 2.00 0.500 0.707 0.999 0.500 0.707 2.00
95 1.47 2.18 0.478 0.705 1.043 0.457 0.673 2.18
100 1.53 2.38 0.456 0.699 1.087 0.414 0.633 2.38
105 1.59 2.60 0.434 0.689 1.131 0.371 0.589 2.60
110 1.64 2.85 0.412 0.675 1.175 0.329 0.540 2.85
115 1.69 3.14 0.389 0.656 1.221 0.289 0.488 3.14
120 1.73 3.46 0.366 0.634 1.267 0.250 0.434 3.46
125 1.77 3.84 0.343 0.608 1.314 0.214 0.379 3.84
130 1.81 4.28 0.318 0.577 1.363 0.179 0.324 4.28
135 1.85 4.82 0.293 0.542 1.413 0.147 0.271 4.82
140 1.88 5.48 0.267 0.502 1.465 0.117 0.221 5.48
145 1.91 6.33 0.240 0.458 1.519 0.091 0.173 6.33
150 1.93 7.44 0.212 0.409 1.576 0.067 0.130 7.44
Blue numbers indicate that Z-loss factors must be considered.
Welding Innovation Vol. XIX, No. 1, 2002 9

Obtuse angles greater than 100 degrees
For this category, contract drawings should show the
required effective throat. Shop drawings are to show the
required leg dimension, calculated with equation 1, or by
using columns C or D of Table 1 (AWS D1.1-2002, para
2.2.5.2, 2.3.3.2).
Angles of 80–100 degrees
For this group, shop drawings are required to show the
fillet leg size (AWS D1.1-2002, para 2.2.5.2). While not
specifically stated in the code, the assumption is that
contract drawings also show this dimension.
Angles of 60–80 degrees
For this category, contract drawings should show the
required effective throat. Shop drawings are to show the
required leg dimension (AWS D1.1-2002, para 2.2.5.2,
2.3.3.2)
Angles of 30–60 degrees
Contract drawings are to show the effective throat. Shop
drawings are required to “show the required leg dimensions
to satisfy the required effective throat, increased by the Z-
loss allowance ... ” (AWS D1.1-2002, para 2.2.5.2, 2.3.3.3).
The Z-loss factor is used to account for the likely incidence
of poor quality welding in the root of a joint with a small
included angle. The amount of poor quality weld in the root
of the joint is a function of the dihedral angle, the welding
process, and the position of welding. Table 2.2 of D1.1
summarizes this data as contained below:
Table 2. Z-loss dimension.
60 o >Ψ > 45 o FCAW-S 1/8 3 FCAW-S 0 0
FCAW-G 1/8 3 FCAW-G 0 0
Position of Welding Position of Welding
V or OH
H or F
Dihedral Angle Ψ Process Z (in.) Z (mm) Process Z (in.) Z (mm)
SMAW 1/8 3 SMAW 1/8 3
GMAW N/A N/A GMAW 0 0
45 o >Ψ > 30 o
SMAW 1/4 6 SMAW 1/4 6
FCAW-S 1/4 6 FCAW-S 1/8 3
FCAW-G 3/8 10 FCAW-G 1/4 6
GMAW N/A N/A GMAW 1/4 6
Once the Z-loss dimension has been determined, it is added
to the required throat dimension. Even though part of the
weld in the root is considered to be of such poor quality as
to be incapable of transferring stress, the resultant weld will
contain sufficient quality weld metal to permit the transfer of
imposed loads. Figure 4 illustrates this concept.
The data in Table 1 that applies to dihedral angles of
30–60 degrees has been shown in blue numbers because
these values must be modified to account for the Z-loss
factors. Such a modification has not been done for the data
in the Table.
Z
Figure 4. Z-loss.
Acute angles less than 30 degrees
The D1.1 code says that welds in joints with dihedral
angles of less than 30 degrees “shall not be considered as
effective in transmitting applied forces …” and then goes
on to discuss a single exception related to tubular structures.
In that exception, with qualification of the welding
procedure specification, such welds may be used for transferring
applied stresses. For plate (e.g., non-tubular) applications,
such an option is not presented in the code.
The practical application of this principle is that when welds
are placed on the acute side, no capacity is assigned to
the weld. Rather, the full load is assumed to be transferred
with the weld on the obtuse side.
Practical Considerations
The most straightforward, and easiest, approach to determining
the required weld size is to assume two welds with
equal throat dimensions will be used, calculate the
required weld throat dimension, and then calculate the
required fillet weld leg size, using either equation 1 or Table
1, columns A and B. Simple? Yes. Best? Let’s see.
The optimizing method that uses equations 6 and 7 will
result in reduced weld metal volumes. But, reduced how
much? The significance increases with greater rotations
from the 90-degree T-joint orientation. For angles of 80, 70,
and 60 degrees, the differences in weld volume are
approximately 3, 12 and 25%. However, note that these differences
are functions of the leg size squared. Accordingly,
the change in leg size is approximately 1, 6, and 13%. In
practical terms, for dihedral angles between 60 and 120
degrees, there will not be a standard fillet weld leg size
until the welds become quite large. In the case of a 70-
degree dihedral angle, for example, and assuming a 1/8 in.
increment for standard sizing of fillet welds over 1 in. leg
size, the leg size would need to be 2 in. before the optimized
weld size would result in a smaller weld. For a 2 mm
standard size, this would equal a 34 mm fillet.
t
ω
ω
ψ 1
10 Welding Innovation Vol. XIX, No. 1, 2002

For angles less than 60 degrees, there can be significant
differences in weld volume. These are situations where the
Z-loss factor must be considered as well. Thus, for angles
of 30–60 degrees, optimization of weld size makes sense,
and the Z-loss factors can be considered at the same time.
It must be recognized that other code provisions may further
affect these results. For example, when optimized for
minimum weld volume, welds on the obtuse side may be
smaller than minimum fillet weld sizes as contained in Table
5.8 of D1.1. The calculated sizes, if less than these minimums,
must be increased to comply with this requirement.
There does not appear to be any intrinsic value in having
welds on opposite sides of the skewed T-joint be of the
same size. If this approach is used, the resultant weld volumes
will fall somewhere between the results for the same
sized throat and the optimized sizes.
After the welds are detailed, the joint must be welded.
Practical considerations apply here too. It must be recognized
that the ratio of the face width “f” to the throat dimension
“t” constitutes the equivalent of a width-to-depth ratio
for the root pass. On the obtuse side, this ratio is large,
exceeding 1:6 for dihedral angle of 106 degrees or more. It
is very difficult to get a single weld bead to “wash” out this
wide without electrode manipulation (weaving). On the
acute side, the ratio is less than 1:2 for angles of 62
degrees. This can lead to width-to-depth ratio cracking.
Recommendations
When determining fillet weld details for skewed T-joints with
dihedral angles from 60–120 degrees, it rarely matters
which method of proportioning of weld sizes is used. Using
equal throat dimensions is a straightforward method, similar
to what is typically done for fillet welds on either side of a
90-degree T-joint. Unless the weld size is large, optimizing it
will probably not result in a smaller specified weld size.
For fillet welds on skewed T-joints with dihedral angles from
30–60 degrees, the Z-loss factor must be considered.
Based on the specific dihedral angle, the welding process,
and the position of welding, the Z-loss factor can be determined,
and this dimension added to the required weld
throat dimension.
It is important to consider how these dimensions should be
communicated between the designer, fabricator, welder and
inspector. The face dimension is a practical means of verifying
that the proper weld size has been achieved.
Opportunities
Lincoln Electric Technical Programs
Welding Technology Workshop
June 10-14, 2002
July 29 – August 2, 2002
The purpose of this program is to
introduce or enhance knowledge of
current thinking in arc welding safety,
theory, processes, and practices. The
course is designed primarily for welding
instructors, supervisors and professional
welders. Fee: $395.
Welding of Aluminum Alloys,
Theory and Practice
October 15-18, 2002
Designed for engineers, technologists,
technicians and welders who are
already familiar with basic welding
processes, this technical training
program provides equal amounts of
classroom time and hands-on welding.
Fee: $595.
Space is limited, so register early to avoid disappointment. For full details, see
www.lincolnelectric.com/knowledge/training/seminars/
Or call 216/383-2240, or write to Registrar, Professional Programs, The Lincoln Electric
Company, 22801 St. Clair Avenue, Cleveland, OH 44117-1199.
Welding Innovation Vol. XIX, No. 1, 2002 11

The Fall of Skyscrapers
By Henry Petroski
A.S. Vesic Professor of Civil Engineering
and Professor of History
Duke University
Durham, North Carolina
Editor’s Note: This article first
appeared in American Scientist,
Volume 90, January-February 2002.
It is reprinted here with permission.
Copyright, Henry Petroski, 2002.
The article is reprinted as written earlier
this year. In the ensuing months,
many investigations have been performed,
and the level of understanding
of some of the technical aspects of
the World Trade Center collapse have
increased. The results of the FEMA
investigation discussed in this article
are now available at www.house.gov/
science/hot/wtc/wtcreport.htm
The terrorist attacks of September 11,
2001, did more than bring down the
World Trade Center towers. The collapse
of those New York City megastructures,
once the fifth and sixth
tallest buildings in the world, signaled
the beginning of a new era in the planning,
design, construction and use of
skyscrapers . For the foreseeable
future, at least in the West, there are
not likely to be any new super-tall
buildings proposed, and only those
currently under construction will be
added to the skylines of the great
cities of the world. Even the continued
occupancy of signature skyscrapers
may come under scrutiny by their
prime tenants.
Since two hijacked airplanes loaded
with jet fuel were crashed within about
15 minutes of each other into the two
most prominent and symbolic structures
of lower Manhattan, the once
reassuringly low numbers generated
by probabilistic risk assessment seem
irrelevant. What happened in New York
ceased being a hypothetical, incredible
or ignorable scenario. From now on,
12 Welding Innovation Vol. XIX, No. 1, 2002
Figure 1. With the collapse of the World Trade Center towers, the fate of future
skyscraper projects has come into question.
structural engineers must be prepared
to answer harder questions about how
skyscrapers will stand up to the impact
of jumbo jets and, perhaps more
important, how they will fare in the
ensuing conflagration. Architects will
likely have to respond more to questions
about stairwells and evacuation
routes than to those about facades
and spires. Because of the nature of
skyscrapers, neither engineers nor
architects will be able to find answers
that will satisfy everyone.
Inclination, Not Economy
Although the idea of the skyscraper is
modern, the inclination to build upward
is not. The Great Pyramids, with their
broad bases, reached heights unapproached
for the next four millennia.
But even the great Gothic cathedrals,
crafted of bulky stone into an aesthetic
of lightness and slenderness, are
dwarfed by the steel and reinforced
concrete structures of the 20th century.
It was modern building materials
that made the true skyscraper structurally
possible, but it was the mechanical
device of the elevator that made
the skyscraper truly practical.
Ironically, it is also the elevator that
has had so much to do with limiting
the height of most tall buildings to
about 70 or 80 stories. Above that,
elevator shafts occupy more than 25
percent of the volume of a tall building,
and so the economics of renting out
space argues against investing in
greater height.
The World Trade Center towers were
110 stories tall, but even with an elaborate
system of local and express elevators,
the associated sky lobbies and
utilities located in the core still
removed almost 30 percent of the towers’
floor area from the rentable space
category. By all planning estimates,

the World Trade Center towers should
have been viewed as a poor investment
and so might not have been
undertaken as a strictly private enterprise.
In fact, it was the Port of New
York Authority, the bi-state governmental
entity now known as the Port
Authority of New York and New Jersey,
that in the 1960s undertook to build
the towers. With its ability to issue
bonds, the Port Authority could afford
to undertake a financially risky project
that few corporations would dare.
Sometimes private enterprise does
engage in similarly questionable
investments, balancing the tangible
financial risk with the intangible gain in
publicity, with the hope that it will
translate ultimately into profit. This was
the case with the Empire State
Building, completed in 1931 and now
the seventh tallest building in the
world. Although it was not heavily
The Port Authority could
undertake a financially risky
project that few corporations
would dare
occupied at first, the cachet of the
world’s tallest building made it a prestigious
address and added to its realestate
value. The Sears Tower stands
an impressive 110 stories tall, the
same count that the World Trade
Center towers once claimed. This skyscraper
gained for its owner the prestige
of having its corporate name
associated with the tallest building in
the world. The Sears Tower, completed
in 1974, one year after the second
World Trade Center tower was finished,
held that title for more than 20
years-until the twin Petronas Towers
were completed in Kuala Lumpur,
Malaysia, in 1998, emphasizing the
rise of the Far East as the location of
new megastructures.
Building Innovation
It is not only the innovative use of elevators,
marketing and political will that
has enabled super-tall buildings to be
built. A great deal of the cost of such a
structure is in the amount of materials
it contains, so lightening the structure
lowers its cost. Innovative uses of
building materials can also give a skyscraper
more desirable office space.
Now more than 70 years old, the steel
frame of the Empire State Building has
closely spaced columns, which break
up the floor space and limit office layouts.
In contrast, the World Trade
Center employed a tubular-construction
principle, in which closely spaced
steel columns were located around the
periphery of the building. Sixty-footlong
steel trusses spanned between
these columns and the inner structure
of the towers, where further columns
were located, along with the elevator
shafts, stairwells and other non-exclusive
office space. Between the core
and tube proper, the broad columnless
space enabled open, imaginative
and attractive office layouts.
The tubular concept was not totally
new with the World Trade Center, it
having been used in the diagonally
braced and tapered John Hancock
Center, completed in Chicago in 1969.
The Sears Tower is also a tubular
structure, but it consists of nine 75-
foot(23-meter)-square tubes bundled
together at the lower stories. The varying
heights of the tubes give the Sears
Tower an ever-changing look, as it presents
a different profile when viewed
from the different directions from
which one approaches it when driving
the city’s expressways. When new to
the Manhattan skyline, the unrelieved
209-foot(64-meter)-square plans and
unbroken 1,360-foot(415-meter)-high
profiles of the twin World Trade Center
buildings came in for considerable
architectural criticism for their lack of
character. Like the Sears Tower, however,
when viewed from different
angles, the buildings, especially as
they played off against each other,
enjoyed a great aesthetic synergy. The
view of the towers from the walkway of
the Brooklyn Bridge was especially
striking, with the stark twin monoliths
echoing the twin Gothic arches of the
bridge’s towers.
Although the World Trade Center towers
did look like little more than tall
prisms from afar, the play of the everchanging
sunlight on their aluminumclad
columns made them new
buildings by the minute. From a closer
perspective, the multiplicity of unbroken
columns corseting each building
also gave it an architectural texture.
The tubular concept
was not totally new
with the WTC
The close spacing of the columns was
dictated by the desire to make the
structure as nearly a perfect tube as
possible. A true tube, like a straw,
would be unpunctured by peripheral
openings, but since skyscrapers are
inhabited by people, windows are considered
a psychological must. At the
same time, too-large windows in very
tall buildings can give some occupants
an uneasy feeling. The compromise
struck in the World Trade Center was
to use tall but narrow windows
between the steel columns. In fact, the
width of the window openings was
said to be less than the width of a person’s
shoulders, which was intended
by the designers to provide a measure
of reassurance to the occupants.
Since the terrorist attack, however,
one of the most haunting images of
those windows is of so many people
standing sideways in the openings,
clinging to the columns and, ultimately,
falling, jumping or being carried to
their death.
Welding Innovation Vol. XIX, No. 1, 2002 13

Failure Analyses
Terrorists first attempted to bring the
World Trade Center towers down in
1993, when a truck bomb exploded in
the lower-level public garage, at the
base of the north tower. Power was
lost in the tower and smoke rose
through it. It was speculated that the
terrorists were attempting to topple the
north tower into the south one, but
even though several floors of the
garage were blown out, the structure
stood. There was some concern
among engineers then that the basement
columns, no longer braced by
the garage floors, would buckle, and
so they were fitted with steel bracing
before the recovery work proceeded.
After that attack, access to the underground
garage was severely restricted,
and security in the towers was considerably
increased. No doubt the 1993
bombing was on the minds of many
people when the airplanes struck the
towers last September.
As they had in the earlier bombing, the
World Trade Center towers clearly survived
the impact of the Boeing 767 airliners.
Given the proven robustness of
the structures to the earlier bombing
assault, the thought that the buildings
might actually collapse was probably
far from the minds of many of those
The survival of the WTC
after the 1993 bombing
seems to have given
an unwarranted sense
of security
who were working in them on September
11. It certainly appears not to have
been feared by the police and fire
fighters who rushed in to save people
and extinguish the fires. Indeed, the
survival of the World Trade Center
after the 1993 bombing seems to have
given an unwarranted sense of security
that the buildings could withstand
even the inferno created by the estimated
20,000 gallons (76,000 liters) of
jet fuel that each plane carried. (That
amount of fuel has been estimated to
have an energy content equivalent to
about 2.4 million sticks of dynamite.)
Steel buildings are expected to be fireproofed,
and so the World Trade
Center towers were. However, fireproofing
is a misnomer, for it only insulates
the steel from the heat of the fire
for a limited period, which is supposed
to be enough time to allow for the fire
to be brought under control, if not
extinguished entirely. Unfortunately, jet
fuel burns at a much higher temperature
than would a fire fed by normal
construction materials and the customary
furniture and contents found in an
office building. Furthermore, conventional
fire-fighting means, such as
water, have little effect on burning jet
fuel. The World Trade Center fire, estimated
to have produced temperatures
as high as of the order of the melting
point of steel, continued unabated. It
has been speculated that some of the
steel beams and columns of the structure
that were not destroyed by the
impact eventually may have been
heated close to if not beyond their
melting point, but this appears to have
been unlikely.
Even if it did not melt, the prolonged
elevated temperatures caused the
steel to expand, soften, sag, bend and
creep. The intense heat also caused
the concrete floor, no longer adequately
supported by the steel beams and
columns in place before the impact of
the airplane, to crack, spall and break
up, compromising the synergistic
action of the parts of the structure.
Without the stabilizing effect of the stiff
floors, the steel columns still intact
became less and less able to sustain
the load of the building above them.
When the weight of the portion of the
building above became too much for
the locally damaged and softened
structure to withstand, it collapsed
onto the floors below. The impact of
the falling top of the building on the
lower floors, whose steel columns
were also softened by heat transfer
along them, caused them to collapse
in turn, creating an unstoppable chain
reaction. The tower that was struck
second failed first in part because the
plane hit lower, leaving a greater
weight to be supported above the
damaged area. (The collapse of the
lower floors of the towers under the
falling weight of the upper floors
occurred for the same reason that a
book easily supported on a glass table
can break that same table if dropped
on it from a sufficient height.)
Within days of the collapse of the towers,
failure analyses appeared on the
Internet and in engineering classrooms.
Perhaps the most widely circulated
were the mechanics-based
analysis of Zdenek Bazant of
Northwestern University and the energy
approach of Thomas Mackin of the
University of Illinois at Urbana-
Champaign. Each of these estimated
that the falling upper structure of a
World Trade Center tower exerted on
the lower structure a force some 30
times what it had once supported.
Charles Clifton, a New Zealand structural
engineer, argues that the fire was
not the principal cause of the collapse.
He thinks that it was the damaged
core rather than the exterior tube
columns that succumbed first to the
enormous load from above. Once the
core support was lost on the impacted
floors, there was no stopping the progressive
collapse, which was largely
channeled by the structural tube to
occur in a vertical direction. In the
wake of the World Trade Center disaster,
the immediate concerns were, of
course, to rescue as many people as
might have survived. Unfortunately,
even to recover most of the bodies
proved an ultimately futile effort. The
twin towers were gigantic structures.
Each floor of each building encompassed
an acre, and the towers
enclosed 60 million cubic feet each.
Together, they contained 200,000 tons
of steel and 425,000 cubic yards (325
cubic meters) or about 25,000 tons of
concrete. The pile of debris in some
places reached as high as a ten-story
building. A month after the terrorist
14 Welding Innovation Vol. XIX, No. 1, 2002

attack, it was estimated that only 15
percent of the debris had been
removed, and it was estimated that it
would take a year to clear the site.
Forensic Engineering
Among the concerns engineers had
about the clean-up operation was how
the removal of debris might affect the
stability of the ground around the site.
Because the land on which the World
Trade Center was built had been part
of the Hudson River, an innovative
barrier had to be developed at the
time of construction to prevent river
water from flowing into the basement
of the structures. This was done with
the construction of a slurry wall, in
which the water was held back by a
deep trench filled with a mudlike mixture
until a hardened concrete barrier
was in place. The completed structure
provided a watertight enclosure, which
came to be known as the “bathtub”
within which the World Trade Center
was built. The basement floors of the
twin towers acted to stabilize the bathtub,
but these were crushed when the
towers broke up and collapsed into the
enclosure. Early indications were that
the bathtub remained intact, but in
order to be sure its walls do not collapse
when the last of the debris and
thus all the internal support is
removed, vulnerable sections of the
concrete wall were being tied back to
the bedrock under the site even as the
debris removal was proceeding.
Atop the pile of debris, the steel
beams and columns were the largest
and most recognizable parts in the
wreckage. The concrete, sheetrock
and fireproofing that were in the building
were largely pulverized by the collapsing
structure, as evidenced by the
ubiquitous dust present in the aftermath.
(A significant amount of
asbestos was apparently used only in
the lower floors of one of the towers,
bad publicity about the material having
accelerated during the construction of
the World Trade Center. Nevertheless,
in the days after the collapse, the
All of the speculations
about the mechanism
of collapse are in fact
hypotheses
once-intolerant Environmental
Protection Agency declared the air
safe.) The grille-like remains of the
buildings’ facades, towering precariously
over what came to be known as
Ground Zero, became a most eerie
image. Though many argued for leaving
these cathedral wall-like skeletons
standing as memorials to the dead,
they posed a hazard to rescue workers
and were in time torn down and
carted away for possible future reuse
in a reconstructed memorial. As is
often the case following such a
tragedy, there was also some disagreement
about how to treat the
wreckage generally. Early on, there
was clearly a need to remove as much
of it from the site as quickly as possible
so that what survivors there might
be could be uncovered. This necessitated
cutting up steel columns into
sections that could fit on large flatbed
trucks. Even the disposal of the wreckage
presented a problem. Much of the
steel was marked for immediate recycling,
but forensic engineers worried
that valuable clues to exactly how the
structures collapsed would be lost.
All of the speculations of engineers
about the mechanism of the collapse
are in fact hypotheses, theories of
what might have happened. Although
computer models will no doubt be constructed
to test those hypotheses and
theories, actual pieces of the wreckage
may provide the most convincing
confirmation that the collapse of the
structures did in fact progress as
hypothesized. Though the wreckage
may appear to be hopelessly jumbled
and crushed, telltale clues can survive
among the debris. Pieces of partially
melted steel, for example, can provide
the means for establishing how hot the
fire burned and where the collapse
might have initiated. Badly bent
columns can give evidence of buckling
before and during collapse. Even the
scratches and scars on large pieces of
steel can be useful in determining the
sequence of collapse. This will be the
task of teams of experts announced
shortly after the tragedy by the
American Society of Civil Engineers
and the Federal Emergency
Management Agency. Also in the
immediate wake of the collapse, the
National Science Foundation awarded
eight grants to engineering and social
science researchers to assess the
debris as it is being removed and to
study the behavior of emergency
response and management teams.
Analyzing the failure of the towers is a
Herculean task, but it is important that
engineers understand in detail what
happened so that they incorporate the
lessons learned into future design
practices. It was the careful failure
analysis of the bombed Federal
Building in Oklahoma City that led
engineers to delineate guidelines for
designing more terrorist-resistant
buildings. The Pentagon was actually
undergoing retrofitting to make it better
able to withstand an explosion when it
was hit by a third hijacked plane on
September 11. Part of the section of
the building that was struck had in fact
just been strengthened, and it suffered
much less damage than the old section
beside it, thus demonstrating the
effectiveness of the work.
Understanding how the World Trade
Center towers collapsed will enable
engineers to build more attack-resistant
skyscrapers. Even before a
detailed failure analysis is completed,
however, it is evident that one way to
minimize the damage to tall structures
is to prevent airplanes and their fuel
from being able to penetrate deeply
into the buildings in the first place.
This is not an impossible task. When a
B-25 bomber struck the Empire State
Welding Innovation Vol. XIX, No. 1, 2002 15

Building in 1945, its body stuck out
from the 78th and 79th floors like a
long car in a short garage. The building
suffered an 18-by-20-foot (5.5-by-
6-meter) hole in its face, but there was
no conflagration, and there certainly
was no collapse. The greatest damage
was done by the engines coming
loose and flying like missiles through
the building. The wreckage of the
plane was removed, the local damage
repaired and the building restored to
its original state. Among the differences
between the Empire State
Building and World Trade Center incidents
was that in the former case, relatively
speaking, a lighter plane struck
a heavier structure. Furthermore, the
propeller-driven bomber was on a
short-range flight from Bedford,
Massachusetts to Newark airport and
The towers might have stood
after the attack if the fires had
been extinguished quickly
so did not have on board the amount
of fuel necessary to complete a
transcontinental flight or to bring down
a skyscraper.
Modern tall buildings can be strengthened
to be more resistant to full penetration
by even the heaviest of aircraft.
This can be done by placing more and
heavier columns around the periphery
of the structure, making the tube
denser and thicker, as it were. The ultimate
defense would be to make the
facade a solid wall of steel or concrete,
or both. This would eliminate windows
entirely, of course, which would defeat
some of the purpose of a skyscraper,
which is in part to provide a dramatic
view from a prestigious office or board
room. The elimination of that attraction,
in conjunction with the increased mass
of the structure itself, would provide
space that would command a significantly
lower rent and yet cost a great
deal more to build. Indeed, no one
would likely even consider building or
Courtesy of American Scientist
Figure 2. Structural design of the
World Trade Center towers was a tube
with a central core.
renting space in such a building.
Hence, the solution would be a Pyrrhic
victory over terrorists.
The World Trade Center towers might
have stood after the terrorist attack if
the fires had been extinguished quickly.
But even if the conventional sprinkler
systems had not been damaged,
water would not have been effective
against the burning jet fuel. Perhaps
skyscrapers could be fitted with a
robust fire-fighting system employing
the kind of foam that is laid down on
airport runways during emergency
landings, or fitted with some other oxygen-depriving
scheme, if there could
also be a way for fleeing people to
breathe in such an environment. Such
schemes would need robustness and
redundancy to survive tremendous
impact forces, so any such system
might be unattractively bulky and prohibitively
expensive to install. Other
approaches might include more effective
fireproofing, such as employing
ceramic-based materials, thus at least
giving the occupants of a burning
building more time to evacuate.
The evacuation of tall buildings will no
doubt now be given much more attention
by architects and engineers alike.
Each World Trade Center tower had
multiple stairways, but all were in the
single central core of the building. In
contrast, stairways in Germany, for
example, are required to be in different
corners of the building. In that configuration,
it is much more likely that one
stairway will remain open even if a
plane crashes into another corner. But
locating stairwells in the corners of a
building means, of course, that prime
office space cannot be located there.
In other words, most measures to
make buildings safer also make them
more expensive to build and diminish
the appeal of their office space. This
dilemma is at the heart of the reason
why the future of the skyscraper is
threatened.
It is likely that, in the wake of the
World Trade Center collapses, any
super-tall building currently in the
development stage will be put on hold
and reconsidered. Real-estate
investors will want to know how the
proposed building will stand up to the
crash of a fully fueled jumbo jet, how
hot the ensuing fire will burn, how long
it will take to be extinguished and how
long the building will stand so that the
occupants can evacuate. The investors
will also want to know who will rent the
space if it is built.
Potential tenants will have the same
questions about terrorist attacks.
Companies will also wonder if their
employees will be willing to work on
the upper stories of a tall building.
Managers will wonder if those employees
who do agree to work in the building
will be constantly distracted,
watching out the window for approaching
airplanes. Corporations will wonder
if clients will be reluctant to come to a
place of business perceived to be vulnerable
to attack. The very need to
have workers grouped together on
adjacent floors in tall buildings is also
being called into question.
After the events of September 11, the
incentive to build a signature structure,
a distinctive super-tall building that
sticks out in the skyline, is greatly
diminished. In the immediate future, as
leases come up for renewal in existing
skyscrapers, real-estate investors will
be watching closely for trends. It is
unlikely that our most familiar skylines
will be greatly changed in the foreseeable
future. Indeed, if companies begin
16 Welding Innovation Vol. XIX, No. 1, 2002

to move their operations wholesale out
of the most distinctive and iconic of
super-tall buildings and into more nondescript
structures of moderate height,
it is not unimaginable that cities like
New York and Chicago will in time see
the reversal of a long-standing trend.
We might expect no longer to see
developers buying up land, demolishing
the low-rise buildings on it, and
putting up a new skyscraper. Instead,
owners might be more likely to demolish
a vacant skyscraper and erect in
its place a building that is not significantly
smaller or taller than its neighbors.
Skylines that were once
immediately recognizable even in silhouette
for their peaks and valleys
may someday be as flat as a mesa.
There is no imperative to such an
interplay between technology and
society. What really happens in the
coming years will depend largely on
how businesses, governments and
individuals react to terrorism and the
threat of terrorism. Unfortunately, the
image of the World Trade Center towers
collapsing will remain in our collective
consciousness for a few
Skylines once recognizable in
silhouette may someday be as
flat as a mesa
generations, at least. Thus, it is no idle
speculation to think that it will be at
least a generation before skyscrapers
return to ascendancy, if they ever do.
Developments in micro-miniaturization,
telecommunications, information technology,
business practice, management
science, economics, psychology
and politics will likely play a much larger
role than architecture and engineering
in determining the immediate
future of macro-structures, at least in
the West.
Bibliography
Bazant, Zdenek P., and Youg Zhou. 2001. “Why
did the World Trade Center collapse?-Simple
analysis.” Journal of Engineering Mechanics,
vol. 128 (2002) pp 2-6. www3.tam.uiuc.edu/
news/200109 c/
Clifton, G. Charles. 2001. Collapse of the World
Trade Center towers. www.hera.org.nz/pdf
files/worldtrade centre.pdf
Mackin, Thomas J. 2001. Engineering analysis
of tragedy at WTC. Presentation slides for ME
346, Department of Mechanical Engineering,
University of Illinois at Urbana-Champaign.
Opportunities
Lincoln Electric Professional Programs
Blodgett’s Design of Welded
Structures
September 24-26, 2002
Blodgett’s Design of Steel Structures is
an intensive 3-day program which
addresses methods of reducing costs,
improving appearance and function,
and conserving material through the
efficient use of welded steel in a broad
range of structural applications.
Seminar leaders: Omer W. Blodgett and
Duane K. Miller. 2.0 CEUs. Fee: $595.
Blodgett’s Design of Weldments
June 4-6, 2002
October 29-31, 2002
Blodgett’s Design of Steel Weldments
is an intensive 3-day program for those
concerned with manufacturing machine
tools, construction, transportation,
material handling, and agricultural
equipment, as well as manufactured
metal products of all types. Seminar
leaders: Omer W. Blodgett and Duane
K. Miller. 2.0 CEUs. Fee: $595.
Fracture & Fatigue Control
in Structures:
Applications of Fracture Mechanics
October 15-17, 2002
Fracture mechanics has become the
primary approach to analyzing and
controlling brittle fractures and fatigue
failures in structures. This course will
focus on engineering applications
using actual case studies. Guest seminar
leaders: Dr. John Barsom and Dr.
Stan Rolfe. 2.0 CEUs. Fee: $595.
Space is limited, so register early to
avoid disappointment. For full details, see
http://www.lincolnelectric.com/
knowledge/training/seminars/
Or call 216/383-2240, or write to
Registrar, Professional Programs, The
Lincoln Electric Company, 22801 St. Clair
Avenue, Cleveland, OH 44117-1199.
Welding Innovation Vol. XIX, No. 1, 2002 17

Lessons Learned in the Field
By Omer W. Blodgett, Sc.D., P.E.
Consider the Transfer of Stress through Members
Introduction
Here I am, back again. In the second
issue of 2001 (Vol. XVIII, No. 2), the
editors of Welding Innovation were
delighted to publish an excellent piece
in this space: “Persistence Pays Off” by
Rob Lawrence of Butler Manufacturing.
Now, where are the submissions from
the rest of you out there?
I started this column a couple of years
ago with the idea of providing a forum
in which our readers could share the
important principles gleaned from the
everyday challenges of working in the
field. It seems to me that often the
“evident” solution to a problem turns
out to be a dead end. I call these my
“ah-ha!” moments. Surely they’ve happened
to many of you. Think about
what you actually learned from these
experiences, that you were able to
apply again in other situations. Then
send an email describing your column
idea to Assistant Editor Jeff Nadzam
at Jeffrey_Nadzam@lincolnelectric.com.
Don’t worry about preparing a finished,
illustrated article. Our writers,
editors and artists can help with that.
We’re just looking for a description of
the real-life circumstances, and a
statement of what you learned.
All right, then, here are some more
lessons I learned, not in school, but
working in the field.
Provide a Path for
Transfer of Stress
A common design oversight is the failure
to provide a path so that a transverse
force can enter that part of the member
(section) that lies parallel to the force.
18 Welding Innovation Vol. XIX, No. 1, 2002
Figure 1.
Given what is needed for the proper
transfer of force (as shown in Figure 1),
let’s consider some examples.
Figure 2.
Figure 3.
The top of Figure 2 shows a lug that
has been welded to a flanged beam in
the simplest and most efficient manner—so
the force goes into the web,
the part parallel to it. In the center
sketch of Figure 2, the lug is placed
across the bottom flange, necessitating
the use of either rectangular or triangular
stiffeners to transfer the load
to the web. If, for some reason, the circumstances
require the lug to be
placed in this manner, the stiffeners
(with the attendant increase in welding
and material usage they entail) are
mandatory. Merely welding the lug
across the more flexible flange could
result in an uneven load on the weld.
Note that the stiffeners are not welded
to the top flange. There would be no
reason to weld them there, since the
flange will not take the force. At the
bottom of Figure 2, the member is in a
different position, and the lug is correctly
welded to the flanges that will
take the load. It is not welded to the
web, since that would serve little purpose
in transferring the force.
Figure 3 illustrates how a lug might be
welded to a box section so as to transfer
force to the parts parallel to it. The
sketch at the top, of course, is not
applicable to the rolled section shown,
since there would be no way of getting
the diaphragm inside the box. But if it
were a fabricated box section, the
diaphragm could be welded in before
welding the top plate on. The center

and bottom drawings in Figure 3 show
additional ways to attach a lug to a
box section. In the center, the lug is
shaped as a sling and directly welded
to the flange. At the bottom, the lug is
designed so it will transfer the force
into the two webs. This is a very efficient
way to transfer the force on the
lug into the webs.
When the Member Is Circular
Figure 4 illustrates two methods of
applying a transverse force to a circular
member. The rationale for these
methods of attachment is shown in
Figure 4.
Figure 5. At the top of Figure 5, the
beam is welded to a support. In standard
practice, it is assumed that the
flanges transfer the bending moments
and the web transfers vertical shear. In
the case of the circular member at the
Figure 5.
bottom of Figure 5, however, it is difficult
to decide which part of the member
is flange, and which part is web.
Mathematical analysis has shown that
if a tube is divided into four quadrants,
the top and bottom quadrants will
transfer 82% of the bending moment,
and the side quadrants, 82% of the
vertical shear. The methods of attaching
the lug shown in Figure 4, therefore,
are methods that transfer force
tending to cause vertical shear into the
areas of the circular section most
closely parallel to the force.
More Complicated Examples
Figure 6 provides a more complicated
example of force transfer. A tank to
haul water on a truck is made up of
1/4 in. (6.4 mm) thick plate, with the
sides overlapping the ends so as to
provide fillet welds. Considering the
forces from the water pressure on the
tank ends, the only place for them to
Figure 6.
go is through the welds and into the
sides—the parts parallel to their direction.
The forces get there by bending
the end plate. In service, the welds
cracked. Three remedies were tried
successively, as shown in Figure 6,
using longitudinal and corner stiffeners,
and finally both longitudinal and
end stiffeners with corner stiffeners.
Figure 7 shows the center sill of a
piggyback railroad car to which a
bracket is welded to carry a 500 lb.
(227 kg) air compressor unit. There are
no interior diaphragms. The vertical
Figure 7.
force from the weight of the unit is
transferred as moment into the bracket,
creating bending at the web. The two
horizontal bending forces must eventually
transfer to the parallel flanges, but
with an open box section there are no
ready pathways. As a result, the web
flexes and fatigue cracks appear in the
web. The sketches at the bottom of
Figure 7 illustrate two possible means
for correcting the faulty design. In one,
a stiffener is added before the web
opposite the bracket side is welded
into the assembly. The stiffener is welded
to both flanges and to one web.
There are now paths for the bending
forces to get to the flanges. The second
way to correct the design is to
shape the bracket so it can be welded
directly to the sill flanges in new fabrications,
or to add pieces to the bracket
on existing cars to accomplish the
same purpose.
Conclusion
The foregoing are just a few examples
intended to illustrate the importance of
considering the transfer of force
through members. Sometimes we
engineers act a little like horses with
blinders on: we concentrate so singlemindedly
on the problem at hand, that
we can’t see what is going on around
us. The ideas discussed in this column
should demonstrate how critical it is
for us as engineers to take our blinders
off, expand our limited views, and
test our assumptions.
Welding Innovation Vol. XIX, No. 1, 2002 19

Gleaming “Waterfall” Refreshes Urban Campus
By Carla Rautenberg
Welding Innovation Contributing Writer
The James F. Lincoln Arc Welding Foundation
Cleveland, Ohio
Controversy and intense interest have
surrounded The Peter B. Lewis
Building of the Weatherhead School of
Management at Case Western
Reserve University ever since architect
Frank Gehry unveiled his design. But
many Cleveland area residents who
shuddered when they first saw photographs
of the model in the city’s
paper, The Plain Dealer, have been
won over as they watched the shining
sculptural curves of the roof take solid
shape and form.
Figure 1.
Figure 2.
According to Plain Dealer architecture
critic Steven Litt, the building depicts
“Gehry’s vision of a gleaming waterfall
splashing over boulders in a mountain
stream.” Sure enough, the stainless
steel skin that slinks over and around
the sensuous curves of the steel structure
glistens in the sun like so much
rushing water. The sight, etched against
a blue sky, can be breathtaking.
What no longer shows is the meticulous
planning, shop fabricating and
field welding work that went into creating
the structural steel supports for
that elegant silvery “gown” of shingles
the $61.7 million building now wears.
When the project’s lead contractor,
Hunt Construction Group of
Indianapolis, called for bids to fabricate
and erect the structural steel, the
response was apparently less than
overwhelming. However, Mariani Metal
Fabricators, Ltd., based across Lake
Erie from Cleveland in Toronto,
Ontario, answered the call.
Software Jumps Industries
Greg Kern, vice president of the 16-
year-old firm, readily admits that the
project was a challenge, not only to
build, but to price. “This was one of the
first uses of CATIA software in the
steel construction industry,” he points
out. CATIA, which was developed for
automotive design and drafting applications,
is employed by architect Frank
Gehry. “Therefore, the geometry was
there for us, because [Gehry] had
worked out the models,” says Kern.
After winning the $6 million structural
steel contract, Mariani Metal hired a
drafting and software training company
with automotive industry expertise to
produce project models and about
1,200 drawings using CATIA. Parallel
CATIA software stations were established
in the Mariani fabrication shop
and at the construction site to provide
design and fabrication adjustments in
real-time, a step which prevented many
potential disruptions in production.
Devising a Practical Approach
When considered in the light of traditional
steel construction concepts, creating
the three-dimensional negative
and positive curves that comprise the
roof structure posed practical problems
both structurally and in terms of
cost. After analyzing the complex
geometry from a real-world erection
standpoint, Mariani Metal proposed
fabricating the structural framework in
a series of ladders, infills, truss panels
and support members, which would be
shop-fabricated (Figure 1) and then
assembled and field-welded on site. “It
was basically a modular approach,”
notes Kern.
Figure 3.
20 Welding Innovation Vol. XIX, No. 1, 2002

Another challenge was devising a
method of bending the hundreds of
pipes which form the “lines of ruling”
(Figures 2 and 3). Almost every pipe
utilized in the roof structure had a
unique curvature and length. With the
help of the CATIA software, Mariani
Metal developed a system which permitted
unique members to be created
within a production line. This allowed
stockpiling of pipe sections which
could then efficiently feed the shop
fabrication process.
Standard AWS D1.1 connection
details were employed to weld the 700
tons of Grade 50 pipe and structural
steel used to create the framework for
the roof. Mariani Metals, which has a
full-time workforce of thirty, employed
eight welders on the job in the shop;
field welding was done by a crew that
averaged between eight and sixteen
welders. The process most used was
shielded metal arc, with semi-automatic
flux cored arc welding in selected
applications. “We kept the field welding
operation as straight-forward as possible,
doing the most complex welding in
the shop,” says Kern. He proudly
states that “the skin lies directly on our
pipes, with the pipes themselves creating
the geometry of the surface,” and
adds that the whole process required
the precision of “building a Swiss
watch in full scale.” That kind of precision
is apparently something the folks
at Mariani thrive on; Kern maintains
that they would hasten to work on
additional Gehry projects.
Across the street from the construction
site in Cleveland’s University Circle,
the stone caryatids of Case Western
Reserve’s gothic style Mather
Memorial Building have silently
watched a monument to 21st century
architecture take dramatic form. When
the 149,000 ft. 2 (13,843 m 2 ) Peter B.
Lewis Building is dedicated later this
year, the city of Cleveland will have
a new landmark.
This soaring structural steel framework is now hidden by the outer skin and
inner walls of the building.
Welding Innovation Vol. XIX, No. 1, 2002 21

P.O. Box 17188
Cleveland, OH 44117-1199
The
James F. Lincoln
Arc Welding
Foundation
NON-PROFIT ORG.
U.S. POSTAGE
PAID
JAMES F. LINCOLN
ARC WELDING FND.
Story on page 20.